There is a need to develop high-efficiency protein production systems using recombinant microorganisms to analyze human genome sequence data recently obtained through the Human Genome Project and functions of diverse proteins identified at genome units and to produce protein products important in human medical fields. When an expression system is selected to produce a recombinant protein derived from higher organisms such as humans, a variety of factors should be carefully considered, which include growth characteristics of host cells, protein expression levels, possibility of intracellular and extracellular expression, possibility of posttranslational modification, biological activity of expressed proteins. As representative microbial expression systems, E. coli and yeast systems are mainly used. E. coli is advantageous because many E. coli-based expression systems have been developed and E. coli expresses heterologous proteins in high levels. However, E. coli has the following drawbacks: the inability to perform posttranslational modification for recombinant production of proteins derived from higher eukaryotes, the difficulty in complete secretion of proteins into the culture medium, the lack of folding ability of proteins possessing many disulfide bonds, and the expression of proteins in insoluble forms such as inclusion bodies (Makrides, Microbial Rev., 1996, 60, 512). In addition, since medically valuable disease-associated proteins among human proteins are mostly glycoproteins or membrane proteins, they need glycosylation and folding into a correct three-dimensional structure through disulfide bonds in order to achieve full activity. Thus, these proteins are impossible to produce in E. coli and essentially require eukaryotic expression systems such as yeasts.
Yeast Saccharomyces cerevisiae is a eukaryotic microorganism proven to be safe to the human body as a GRAS (Generally Recognized As Safe) organism. S. cerevisiae has many advantages including easy gene manipulation, various developed expression systems and easy large-scale culture. The advantages further include that S. cerevisiae functions to secrete higher cell-derived proteins such as human proteins into the extracellular space, and performs posttranslational modification of proteins, such as glycosylation. The extracellular secretion can be achieved through the artificial fusion of a target protein with a protein secretory signal, and during the secretion of a protein, protein folding or disulfide bond formation and glycosylation occur, thereby producing a fully biologically active recombinant protein. Also, since a biologically active protein can be obtained directly from the culture medium, S. cerevisiae-based protein expression systems do not require cost-inefficient cell disruption or refolding so that they are very economical (Eckart and Bussineau, Curr. Opin. Biotechnol., 1996, 7, 525).
However, despite the many advantages of S. cerevisiae mentioned above, the problem of present techniques associated with systems for secreting human proteins using yeast S. cerevisiae involves non-uniform protein secretion yield ranging from no production to several grams/liter, depending on the human protein, leading to a great difference of more than several thousands in protein secretion yield, thus making it difficult to predict secretion yield. When a heterologous protein is secreted in several grams/liter, this protein production is considered to be cost-effective. In contrast, for the production of proteins expressed in low levels, especially highly valuable human therapeutic proteins, difficulties often occur in the expression and secretion of the proteins. To solve these problems, much research has been focused on secretory factors involved in protein secretion. For example, many studies have been carried out on chaperons, including a method of overexpressing a secretory factor, BiP (KAR2), which helps fold proteins newly synthesized in the endoplasmic reticulum(ER) (Robinson et al., Biotechnol. prog., 1996, 271, 10017), and a method of overexpressing PDI (protein disulfide isomerase) helping the formation of cysteine bonds (Robinson et al., Bio/Technology, 1994, 12, 381; Schultz et al., Ann. N. Y. Acad. Sci., 1994, 721, 148; Hayano et al., FEBS Lett., 1995, 377, 505). Also, another study has been performed to improve secretion through preparation of a fusion partner inducing secretion and fusion with a well-secreted protein (Gouka et al., Appl. Microbiol. Biotechnol., 1997, 47, 1). To date, these methods have been considered to be very successful in improving the secretion of heterologous proteins. Molecular mechanisms of these fusion techniques have been poorly studied, but these fusion techniques have been experimentally proven to reduce limitations in translational or posttranslational steps, including facilitating protein translocation and helping protein folding.
Kjeldsen et al. (Protein Expr. Purif., 1997, 9, 331) enhanced the secretion of insulin by fusing insulin precursor with a synthetic leader prepared based on theoretical consideration in order to achieve effective secretion of insulin or insulin precursor. The synthetic leader has an N-glycosylation site and a BiP recognition site so that it extends the residence of the fusion protein in the ER, leading to correct folding of the insulin precursor. Also, the synthetic leader in which an additional glycosylation site is introduced remarkably increased the secretion of insulin in Aspergillus niger and Saccharomyces cerevisiae (Kjeldsen et al., Protein Expr. Purif., 1998, 14, 309). Similar results were obtained in Aspergillus awamori (Ward et al., Bio/Technology, 1989, 8, 435) and when hydrophobic cutinase is expressed in yeast (Sagt et al., Appl. Environ. Microbiol. 2000, 66, 4940). This high-yield secretion of recombinant proteins results from the introduction of glycosylation sites that increase the solubility of recombinant proteins in the ER and induce correct folding of the proteins.
Well-secreted proteins have been employed as fusion partners. For example, fusion expression with glucoamylase from Aspergillus awamori resulted in an increase in secretion yield of the following proteins: bovine prochymosin (Ward et al., Bio/Technology, 1989, 8, 435), porcine pancreatic phospholipase A2 (Roberts et al., Gene, 1992, 122, 155), human interleukin-6 (Contreras et al., Bio/Technology 1991, 9, 378; Broekhuijsen et al., J. Biotechnol., 1993, 31, 135), hen egg-white lysozyme (Jeenes et al., FEMS Microbiol Lett, 1993, 107, 267), and human lactoferrin (Ward et al., Bio/Technology, 1995, 13, 498). Increased secretion yield varied, depending on the protein, in a range of 5 to 1000 times. Also, the use of amino-terminal 24 amino acids of human interleukin-1β as a fusion partner in yeast resulted in an increase in secretion yield of human growth hormone and granulocyte colony-stimulating factor (G-CSF) (Lee et al., Biotechnol. Prog., 1999, 15, 884). Human interleukin-1β is secreted without a particular secretory signal (Muesch et al., Trends Biochem. Sci., 1990, 15, 86), and its recombination production is very effective via secretion in yeast (Baldari et al., Protein Eng., 1987, 1, 433). Also, according to a recent report, a fusion partner originally retained in a protein is essential for correct folding of the protein (Takahashi et al., Appl Microbiol. Biotechnol., 2001, 55, 454). When the mature form of Rhizopus oryzae lipase (ROL) fused to the pre-pro-leader sequence of the mating factor alpha from S. cerevisiae was expressed in order to express ROL in S. cerevisiae, secretion of ROL was not observed. However, when ROL was synthesized together with the prosequence, ROL was properly secreted. These results demonstrate that the prosequence of ROL is essential for the folding of ROL itself.
As described above, through much research, various seretory factors have been developed to induce the secretion of recombinant proteins. However, although the developed secretory factors are effective to increase the secretion level of particular proteins, they cannot be used as a general means for the secretory production of all proteins. Dorner et al. reported that overexpression of BiP in CHO cells rather reduces protein secretion (Dorner et al., EMBO J., 1992, 11, 1563), and decreased BiP expression increases protein secretion (Dorner et al., Mol. cell. Biol., 1988, 8, 4063). In yeast, overexpression of KAR2 (BiP) did not enhance the secretion of plant thaumatin (Harmsen et al., Appl. Microbiol. Biotechnol., 1996, 46, 365). Overexpression of BiP in Baculovirus resulted in an increase in levels of a soluble antibody in cell lysates but did not increase secretion yield of the antibody (Hsu et al., Protein Expr. Purif., 1994, 5, 595). When another secretory factor PDI as a foldase was overexpressed in Aspergillus niger, secretion of glucoamylase did not increase (Wang and Ward, Curr. Genet. 2000, 37, 57). Secretion improvement using a protein fusion partner was also reported to have a problem of increasing the secretion efficiency only of particular proteins.